Patient factors that matter in predicting spine surgery outcomes: a machine learning approach

Joel A. Finkelstein MD, MSc, FRCS(C)1,2, Roland B. Stark MEd3, James Lee MD2, and Carolyn E. Schwartz ScD3,5
View More View Less
  • 1 Divisions of Orthopedic Surgery and
  • | 2 Spine Surgery, Sunnybrook Health Sciences Centre, Toronto, Ontario, Canada;
  • | 3 DeltaQuest Foundation, Inc., Concord, Massachusetts; and
  • | 4 Departments of Medicine and
  • | 5 Orthopaedic Surgery, Tufts University School of Medicine, Boston, Massachusetts
Restricted access

Purchase Now

USD  $45.00

Spine - 1 year subscription bundle (Individuals Only)

USD  $369.00

JNS + Pediatrics + Spine - 1 year subscription bundle (Individuals Only)

USD  $600.00
Print or Print + Online

OBJECTIVE

There is an increasing recognition of the importance of predictive analytics in spine surgery. This, along with the addition of personalized treatment, can optimize treatment outcomes. The goal of this study was to examine the value of clinical, demographic, expectation, and cognitive appraisal variables in predicting outcomes after surgery.

METHODS

This prospective longitudinal cohort study followed adult patients undergoing spinal decompression and/or fusion surgery for degenerative spinal conditions. The authors focused on predicting the numeric rating scale (NRS) for pain, based on past research finding it to be the most responsive of the spine patient-reported outcomes. Clinical data included type of surgery, adverse events, comorbidities, and use of pain medications. Demographics included age, sex, employment status, education, and smoking status. Data on expectations related to pain relief, ability to do household and exercise/recreational activities without pain, preventing future disability, and sleeping comfort. Appraisal items addressed 22 cognitive processes related to quality of life (QOL). LASSO (least absolute shrinkage and selection operator) and bootstrapping tested predictors hierarchically to determine effective predictive subsets at approximately 10 months postsurgery, based on data either at baseline (model 1) or at approximately 3 months (model 2).

RESULTS

The sample included 122 patients (mean age 61 years, with 53% being female). For model 1, analysis revealed better outcomes with patients expecting to be able to exercise or do recreational activities, focusing on recent events, and not focusing on how others see them (mean bootstrapped R2 [R2boot] = 0.12). For model 2, better outcomes were predicted by expecting symptom relief, focusing on the positive and on one’s spinal condition (mean R2boot = 0.38). Bootstrapped analyses documented the stability of parameter estimates despite the small sample.

CONCLUSIONS

Nearly 40% of the variance in spine outcomes was accounted for by cognitive factors, after adjusting for clinical and demographic factors. Different expectations and appraisal processes played a role in long- versus short-range predictions, suggesting that cognitive adaptation is important and relevant to pain relief outcomes after spine surgery. These results underscore the importance of addressing how people think about QOL and surgery outcomes to maximize the benefits of surgery.

ABBREVIATIONS

LASSO = least absolute shrinkage and selection operator; MCID = minimal clinically important difference; NRS = numeric rating scale; NSAID = nonsteroidal anti-inflammatory drug; PRO = patient-reported outcome; QOL = quality of life; SEboot = bootstrapped standard error.

Spine - 1 year subscription bundle (Individuals Only)

USD  $369.00

JNS + Pediatrics + Spine - 1 year subscription bundle (Individuals Only)

USD  $600.00

Contributor Notes

Correspondence Joel A. Finkelstein: Sunnybrook Health Sciences Centre, Toronto, ON, Canada. joel.finkelstein@sunnybrook.ca.

INCLUDE WHEN CITING Published online May 21, 2021; DOI: 10.3171/2020.10.SPINE201354.

Disclosures The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.

  • 1

    Schwartz CE, Ayandeh A, Finkelstein JA. When patients and surgeons disagree about surgical outcome: investigating patient factors and chart note communication. Health Qual Life Outcomes. 2015;13:161.

    • Search Google Scholar
    • Export Citation
  • 2

    Lauridsen HH, Manniche C, Korsholm L, et al. . What is an acceptable outcome of treatment before it begins? Methodological considerations and implications for patients with chronic low back pain. Eur Spine J. 2009;18(12):18581866.

    • Search Google Scholar
    • Export Citation
  • 3

    Schwartz CE, Finkelstein JA. Understanding inconsistencies in patient-reported outcomes after spine treatment: response shift phenomena. Spine J. 2009;9(12):10391045.

    • Search Google Scholar
    • Export Citation
  • 4

    Schwartz CE, Sajobi TT, Lix LM, et al. . Changing values, changing outcomes: the influence of reprioritization response shift on outcome assessment after spine surgery. Qual Life Res. 2013;22(9):22552264.

    • Search Google Scholar
    • Export Citation
  • 5

    Schwartz CE, Zhang J, Rapkin BD, Finkelstein JA. Reconsidering the minimally important difference: evidence of instability over time and across groups. Spine J. 2019;19(4):726734.

    • Search Google Scholar
    • Export Citation
  • 6

    Sprangers MAG, Schwartz CE. Integrating response shift into health-related quality of life research: a theoretical model. Soc Sci Med. 1999;48(11):15071515.

    • Search Google Scholar
    • Export Citation
  • 7

    Schwartz CE, Sprangers MAG. Methodological approaches for assessing response shift in longitudinal health-related quality-of-life research. Soc Sci Med. 1999;48(11):15311548.

    • Search Google Scholar
    • Export Citation
  • 8

    Finkelstein JA, Razmjou H, Schwartz CE. Response shift and outcome assessment in orthopedic surgery: is there a difference between complete and partial treatment?. J Clin Epidemiol. 2009;62(11):11891190.

    • Search Google Scholar
    • Export Citation
  • 9

    Rapkin BD, Schwartz CE. Toward a theoretical model of quality-of-life appraisal: implications of findings from studies of response shift. Health Qual Life Outcomes. 2004;2(1):14.

    • Search Google Scholar
    • Export Citation
  • 10

    Shmueli G. To explain or to predict?. Stat Sci. 2010;25(3):289310.

  • 11

    Brusko GD, Kolcun JPG, Wang MY. Machine-learning models: the future of predictive analytics in neurosurgery. Neurosurgery. 2018;83(1):E3E4.

    • Search Google Scholar
    • Export Citation
  • 12

    Van Esbroeck A, Rubinfeld I, Hall B, Syed Z. Quantifying surgical complexity with machine learning: looking beyond patient factors to improve surgical models. Surgery. 2014;156(5):10971105.

    • Search Google Scholar
    • Export Citation
  • 13

    Kim JS, Merrill RK, Arvind V, et al. . Examining the ability of artificial neural networks machine learning models to accurately predict complications following posterior lumbar spine fusion. Spine (Phila Pa 1976).2018;43(12):853860.

    • Search Google Scholar
    • Export Citation
  • 14

    Khor S, Lavallee D, Cizik AM, et al. . Development and validation of a prediction model for pain and functional outcomes after lumbar spine surgery. JAMA Surg. 2018;153(7):634642.

    • Search Google Scholar
    • Export Citation
  • 15

    Held U, Burgstaller JM, Wertli MM, et al. . Prognostic function to estimate the probability of meaningful clinical improvement after surgery—results of a prospective multicenter observational cohort study on patients with lumbar spinal stenosis. PLoS One. 2018;13(11):e0207126.

    • Search Google Scholar
    • Export Citation
  • 16

    Koen EA, Antonarakos PD, Katranitsa LT, et al. . Lumbar discectomy: generic outcome measures for specific outcome prediction: a short-term, observational, pilot study. Clin Spine Surg. 2017;30(2):E132E137.

    • Search Google Scholar
    • Export Citation
  • 17

    Micankova Adamova B, Vohanka S, Dusek L, et al. . Prediction of long-term clinical outcome in patients with lumbar spinal stenosis. Eur Spine J. 2012;21(12):26112619.

    • Search Google Scholar
    • Export Citation
  • 18

    Zulman DM, Haverfield MC, Shaw JG, et al. . Practices to foster physician presence and connection with patients in the clinical encounter. JAMA. 2020;323(1):7081.

    • Search Google Scholar
    • Export Citation
  • 19

    Hegarty D, Shorten G. Multivariate prognostic modeling of persistent pain following lumbar discectomy. Pain Physician. 2012;15(5):421434.

    • Search Google Scholar
    • Export Citation
  • 20

    Cobo Soriano J, Sendino Revuelta M, Fabregate Fuente M, et al. . Predictors of outcome after decompressive lumbar surgery and instrumented posterolateral fusion. Eur Spine J. 2010;19(11):18411848.

    • Search Google Scholar
    • Export Citation
  • 21

    Block AR, Gatchel RJ, Deardorff WW, Guyer RD. The Psychology of Spine Surgery. American Psychological Association;2003.

  • 22

    Rapkin BD, Schwartz CE. Advancing quality-of-life research by deepening our understanding of response shift: a unifying theory of appraisal. Qual Life Res. 2019;28(10):26232630.

    • Search Google Scholar
    • Export Citation
  • 23

    Ghogawala Z, Dziura J, Butler WE, et al. . Laminectomy plus fusion versus laminectomy alone for lumbar spondylolisthesis. N Engl J Med. 2016;374(15):14241434.

    • Search Google Scholar
    • Export Citation
  • 24

    Försth P, Ólafsson G, Carlsson T, et al. . A randomized, controlled trial of fusion surgery for lumbar spinal stenosis. N Engl J Med. 2016;374(15):14131423.

    • Search Google Scholar
    • Export Citation
  • 25

    SurveyGizmo. Alchemer; 2019.Accessed November 30, 2020. https://www.alchemer.com

  • 26

    Schwartz CE, Stark RB, Balasuberamaniam P, et al. . Responsiveness of standard spine outcome tools: do they measure up?. J Neurosurg Spine. 2020;33(1):106113.

    • Search Google Scholar
    • Export Citation
  • 27

    Childs JD, Piva SR, Fritz JM. Responsiveness of the numeric pain rating scale in patients with low back pain. Spine (Phila Pa 1976).2005;30(11):13311334.

    • Search Google Scholar
    • Export Citation
  • 28

    Davis JA, Weber RP. The Logic of Causal Order. Vol 55.Sage Publications, Inc;1985.

  • 29

    James G, Witten D, Hastie T, Tibshirani R. An Introduction to Statistical Learning With Applications in R. Springer;2014.

  • 30

    van Buuren S. Groothuis-Oudshoorn. mice: Multivariate imputation by chained equations. J Stat Softw. 2011;45(3):i03.

  • 31

    Graham JW. Missing data analysis: making it work in the real world. Annu Rev Psychol. 2009;60:549576.

  • 32

    Azur MJ, Stuart EA, Frangakis C, Leaf PJ. Multiple imputation by chained equations: what is it and how does it work?. Int J Methods Psychiatr Res. 2011;20(1):4049.

    • Search Google Scholar
    • Export Citation
  • 33

    Khan SI, Hoque ASML. SICE: an improved missing data imputation technique. J Big Data. 2020;7(1):37.

  • 34

    Allison PD. Quantitative Applications in the Social Sciences: Missing Data. Sage Publications, Inc;2002.

  • 35

    Freedman DA. Bootstrapping regression models. Ann Stat. 1981;9(6):12181228.

  • 36

    IBM SPSS Statistics for Windows. Version 26.IBM Corp;2019.

  • 37

    R: A language and environment for statistical computing. R Foundation; 2017.Accessed November 30, 2020. https://www.r-project.org

  • 38

    MacKinnon DP, Krull JL, Lockwood CM. Equivalence of the mediation, confounding and suppression effect. Prev Sci. 2000;1(4):173181.

  • 39

    McGinn LK. Cognitive behavioral therapy of depression: theory, treatment, and empirical status. Am J Psychother. 2000;54(2):257262.

  • 40

    Hettema J, Steele J, Miller WR. Motivational interviewing. Annu Rev Clin Psychol. 2005;1:91111.

  • 41

    Moyers TB, Manuel JK, Ernst D. Motivational Interviewing Treatment Integrity Coding Manual 4.1. 2014.Accessed November 25, 2020. https://casaa.unm.edu/download/MITI4_2.pdf

    • Search Google Scholar
    • Export Citation
  • 42

    Cella D, Yount S, Rothrock N, et al. . The Patient-Reported Outcomes Measurement Information System (PROMIS): progress of an NIH roadmap cooperative group during its first two years. Med Care. 2007;45(5 Suppl 1):S3S11.

    • Search Google Scholar
    • Export Citation

Metrics

All Time Past Year Past 30 Days
Abstract Views 417 417 417
Full Text Views 51 51 51
PDF Downloads 66 66 66
EPUB Downloads 0 0 0